base-200k / train_results.json
nancyH's picture
final pretrained checkpoint at step 200000
33429b1 verified
Raw
History Blame Contribute Delete
198 Bytes
{
"epoch": 20.2,
"train_loss": 2.548693740234375,
"train_runtime": 257533.7655,
"train_samples": 950449,
"train_samples_per_second": 74.553,
"train_steps_per_second": 0.777
}